Frequency-domain full-waveform inversion-based musculoskeletal ultrasound computed tomography

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Computational Physics

Background:

• Full-waveform inversion (FWI) is a powerful ultrasound computed tomography (USCT) tool.
• Time-domain FWI (TDFWI) faces computational challenges.
• Frequency-domain FWI (FDFWI) offers improved computational efficiency for imaging.

Purpose of the Study:

• To develop an effective source estimation method for musculoskeletal FDFWI.
• To address challenges in imaging tissues with high impedance contrast, like bone.
• To enhance the convergence and accuracy of FDFWI in musculoskeletal applications.

Main Methods:

• A water-referenced data calibration method was proposed for source estimation.
• A starting frequency criterion was introduced to prevent local minima and improve convergence.
• Numerical studies were conducted on leg models with varying musculoskeletal structures and lesions.

Main Results:

• The proposed water-referenced calibration method effectively addresses source estimation challenges in the presence of bone.
• The starting frequency criterion facilitated algorithm convergence and avoided cycle-skipping.
• Numerical simulations successfully demonstrated the retrieval of musculoskeletal anatomies and lesions.

Conclusions:

• The study successfully extends FDFWI to musculoskeletal USCT imaging.
• The proposed methods provide an efficient and accurate solution for musculoskeletal imaging.
• This work offers a viable alternative for high-resolution musculoskeletal USCT.